and the Rwandan Gacaca Courts
by
Emma Anne Shewell
Thesis presented in fulfilment of the requirements for the degree of Master of Arts in the Faculty of Arts and Social Sciences at
Stellenbosch University
Supervisor: Dr Cindy Steenekamp March 2020
i DECLARATION
By submitting this thesis electronically, I declare that the entirety of the work contained therein is my own, original work, that I am the sole author thereof (save to the extent explicitly otherwise stated), that reproduction and publication thereof by Stellenbosch University will not infringe any third party rights and that I have not previously in its entirety or in part submitted it for obtaining any qualification.
March 2020
Copyright © 2020 Stellenbosch University All rights reserved
ii ABSTRACT
Within the broad field of conflict resolution, mediation has traditionally been conceptualised as a tool for conflict management or settlement. In other words, in both theory and practice, mediation is assumed to be a process of interest-based negotiation facilitated by a third party with the aim of arriving at a tangible agreement. Whether the agreement be in the form of a ceasefire, partial settlement or signed peace treaty, this outcomes-based approach has traditionally dominated the mediation discourse. In African practice this trend remains, with most mediation efforts being restricted to the elite level in the form of political negotiations between high-profile leaders.
This study aims to extend these assumptions of traditional mediation concerning the conflict resolution potential of the mechanism. Given that mediation brings disputants together in communication that is otherwise unlikely to have taken place, it opens a unique possibility for dialogue and improved mutual understanding. By developing a newer model of mediation, that of ‘process-based’ mediation, this study resituates the mechanism as a tool for holistic conflict transformation. This approach prioritises the process of mediation as an end in itself, as opposed to a means to settlement ends. When formulated according to the five characteristics of the process-based model, mediation projects are argued to engender transformations of conflict at their root cause by improving the quality of relational interaction between disputants. These characteristics are community participation, context-specificity, the use of an Insider-Partial mediator, limited resource pressure, and a relational focus.
Through a comparative case study analysis of the South African Truth and Reconciliation Commission and the Rwandan gacaca courts and using the process-based model as an analytical framework, the study demonstrates the applicability of this model in cases of protracted conflict on the African continent. In comparing the relative manifestation of the model within these cases and their subsequent transformative success, the study provides support for the use of process-based mediation as a conflict response in Africa. The case comparison finds that the five characteristics of the process-based model are in many ways linked, with the most important transformative elements being community participation and a relational focus. By demonstrating the transformative potential of mediation practice, the findings contribute to a contemporary movement in mediation literature away from its limited traditional conceptualisation.
iii OPSOMMING
In die breë veld van konflikoplossing is bemiddeling tradisioneel gekonseptualiseer as 'n instrument vir konflikhantering of oplossing. Met ander woorde, in beide teorie en praktyk, word aanvaar dat bemiddeling 'n proses van onderhandeling is wat deur 'n derde party gefasiliteer word op grond van belang, met die doel om tot 'n tasbare ooreenkoms te kom. Of die ooreenkoms in die vorm van 'n wapenstilstand, gedeeltelike skikking of ondertekende vredesverdrag bestaan, oorheers hierdie uitkomsgebaseerde benadering tradisioneel die bemiddelingsdiskoers. In Afrika-praktyke bly hierdie tendens voort, met die meeste bemiddelingspogings beperk tot die elite-vlak in die vorm van politieke onderhandelinge tussen hoëprofielleiers.
Hierdie studie het ten doel gestel om hierdie aannames van tradisionele bemiddeling rakende die konflikoplossingspotensiaal van die meganisme uit te brei. Aangesien bemiddeling geskille bymekaar bring vir kommunikasie wat andersins onwaarskynlik sou plaasvind, skep dit 'n unieke moontlikheid vir dialoog en verbeterde wedersydse begrip. Deur die ontwikkeling van 'n nuwer bemiddelingsmodel, naamlik 'proses-gebaseerde' bemiddeling, word hierdie meganisme hervat as 'n instrument vir holistiese konfliktransformasie. Hierdie benadering prioritiseer die bemiddelingsproses as ‘n doel op sigself, in teenstelling met 'n middel tot skikking. Wanneer bemiddelingsprojekte geformuleer word volgens die vyf kenmerke van die proses-gebaseerde model, word daar geredeneer dat bemiddelingsprojekte transformasie van konflik tot gevolg het by hulle kernoorsaak, deur die kwaliteit van die verhoudingsinteraksie tussen disputante te verbeter. Hierdie kenmerke is gemeenskapsdeelname, konteks-spesifisiteit, die gebruik van 'n “Insider-Partial” bemiddelaar, beperkte hulpbrondruk en 'n verhoudingsfokus.
Deur middel van 'n vergelykende gevallestudie-analise van die Suid-Afrikaanse Waarheids- en Versoeningskommissie en die Rwandese gacaca-howe, en die gebruik van die prosesgebaseerde model as analitiese raamwerk, toon die studie die toepaslikheid van hierdie model in gevalle van uitgerekte konflik op die vasteland van Afrika aan. In die vergelyking van die relatiewe manifestasie van die model in hierdie gevalle en die daaropvolgende transformatiewe sukses, bied die studie ondersteuning vir die gebruik van prosesgebaseerde bemiddeling as konflikrespons in Afrika. Die gevalvergelyking vind dat die vyf eienskappe van die prosesgebaseerde model op baie maniere gekoppel is met die belangrikste transformerende elemente, gemeenskapsdeelname en 'n verhoudingsfokus. Deur die transformatiewe potensiaal van bemiddelingspraktyk te demonstreer, dra die bevindings by tot 'n kontemporêre beweging in bemiddelingsliteratuur en beweeg weg van die beperkte tradisionele konseptualisering.
iv ACKNOWLEDGEMENTS
17 years of education later and it has all come down to this – 135 pages of coffee, tears and a whole heap of pride!
First and foremost, to my supervisor Dr Steenekamp I give my wholehearted thanks. Without your guidance this quite literally would not have been possible. Your meticulous organisation kept me on track and helped me achieve this momentous task in half the time I expected it would. Besides your excellent academic support, your encouragement and good nature always meant I left your office with calmed nerves and a smile on my face. I hope that given my newfound enjoyment of all things SPSS we will one day be able to work together again. Next, to the Shewell Clan, can you believe we’re finally here? After all these years of “Stop stressing Emz” and “It’ll all be okay Emz”, we have made it to the end. Mom, my life-long supervisor, cheerleader and psychologist, through every step of this journey you have been an absolute blessing and I could not have done it without you. Dad, your unwavering belief in me and my abilities helped me to believe in myself even on the most difficult of days. And Dan, having you here this year has been more of a comfort than you know, and I’m so grateful we got to share this special little town for at least one year before going our separate ways again. To my Saltmates, Phili, Tals and Lau, your endless support and celebration of small victories kept me going when it felt like the mountain was becoming too steep to climb. To my roomie Ash, thank you for the coffees poured, the movies watched, and the dishes washed when my to-do list was stacked as high as the plates next to the sink. And to the rest of my girl gang, Shan, Ali, Jen and Mel, thank you for the phone calls, wine dates, and years upon years of friendship. I appreciate you all more than words can express.
And finally, to my beautiful South Africa, thank you for inspiring me to pursue a field of study that sets my heart alight. We may have a ways to go before we can truly claim our Rainbow Nation status, but I have no doubts that the future is oh so bright and colourful. I cannot wait to see what is in store for this magnificent country I am privileged to call home.
v TABLE OF CONTENTS
CHAPTER ONE – RESEARCH OUTLINE ... 1
1.1Introduction ... 1
1.2Background and Rationale of the Study ... 2
1.3Problem Statement and Research Questions ... 4
1.4Research Design and Methodology ... 5
1.4.1Case Study Design ... 5
1.4.2Comparative Analysis Methodology ... 8
1.4.3Alexander George’s Structured, Focused Comparison ... 10
1.4.4Case Selection ... 11
1.4.5Data Collection ... 13
1.4.6Limitations and Challenges ... 14
1.5Contribution of the Study ... 14
1.6Outline of the Study ... 15
CHAPTER TWO – CONCEPTUAL, THEORETICAL AND ANALYTICAL FRAMEWORK ... 17
2.1Introduction ... 17
2.2Mediation Defined ... 17
2.2.1Conflicts Conducive to Mediation Intervention ... 18
2.2.2Bias, Neutrality and Impartiality ... 20
2.2.3Mediator Role and the Use of Leverage ... 22
2.2.4Mediator Identity ... 25
2.2.5Major Threats to Effective Mediation ... 27
2.2.6Difficulties in Measuring Mediation Success ... 28
2.3Questioning the Durability of Mediated Peace ... 29
2.4Short-Term Agreement or Long-Lasting Peace? ... 31
2.4.1A Shift in Theory – Management to Transformation ... 31
2.4.2Resolution vs. Transformation ... 35
2.4.3Concepts of Transformation and Transitional Societies ... 36
2.4.3.1 Reconciliation ... 37
2.4.3.2 Justice ... 39
2.4.3.3 Truth ... 40
2.4.3.4 Peace ... 41
2.4.4Theories of Transformative Mediation ... 43
2.4.4.1 Bush and Folger’s Transformative Mediation ... 43
2.4.4.2Lederach’s Mediative Capacity... 45
2.5 Mediation as a Context-Specific Process ... 47
vi
2.6 Analytical Framework: A ‘Process-Based’ Mediation Model... 50
2.7Conclusion ... 53
CHAPTER THREE – HISTORICAL CONTEXTUALISATION OF RWANDA AND SOUTH AFRICA ... 55
3.1Introduction ... 55
3.2Gacaca – Rebuilding after Genocide in Rwanda ... 55
3.2.1Hutu vs Tutsi - A Colonial Divide? ... 55
3.2.2The Spring of 1994 ... 57
3.2.3The Gacaca Courts ... 60
3.3The Truth and Reconciliation Commission – Uniting a Separated South Africa ... 64
3.3.1A History of Separatism ... 64
3.3.2The Reality of Apartheid ... 65
3.3.3The Truth and Reconciliation Commission ... 71
3.4Conclusion ... 75
CHAPTER FOUR – APPLICATION OF THE ANALYTICAL MODEL AND COMPARATIVE ANALYSIS ... 76
4.1Introduction ... 76
4.2Application of the Analytical Model to South Africa and Rwanda ... 77
4.2.1 Community Participation ... 77 4.2.1.1 Application ... 78 4.2.1.2 Comparison ... 81 4.2.2 Context-specificity... 83 4.2.2.1 Application ... 83 4.2.2.2 Comparison ... 86 4.2.3 ‘Insider-Partial’ Mediator ... 88 4.2.3.1 Application ... 89 4.2.3.2 Comparison ... 92 4.2.4 Resource Pressure ... 93 4.2.4.1 Application ... 93 4.2.4.2 Comparison ... 95 4.2.5 Relational Focus ... 96 4.2.5.1 Application ... 96 4.2.5.2 Comparison ... 101 4.3 Transformative Success ... 102 4.4 Conclusion ... 113
CHAPTER FIVE – CONCLUSION ... 115
5.1 Introduction ... 115
vii
5.3 Main Findings ... 117
5.4 Reflection ... 121
5.5 Recommendations for Further Research ... 123
5.6 Concluding Remarks ... 123
viii LIST OF ABBREVIATIONS
ANC African National Congress
AU African Union
AUMSU African Union Mediation Support Unit
AI Amnesty International
MK Umkhonto we Sizwe
NURC National Unity and Reconciliation Commission PAC Pan-Africanist Congress
PFP Progressive Federal Party PRI Penal Reform International
RRB Rwanda Reconciliation Barometer SACP South African Communist Party SADF South African Defence Force
SARB South African Reconciliation Barometer
SPLA/NDA Sudan People’s Liberation Army/National Democratic Alliance TRC Truth and Reconciliation Commission
UDF United Democratic Front
1 CHAPTER ONE – RESEARCH OUTLINE
1.1 Introduction
The history of the African continent, rife with continued conflict and resultant combat, has shown how significant societal divisions can be when left unresolved. Conflict in Africa appears to persist in cyclical patterns, with there being a strong likelihood of recurrence in countries that have already experienced large-scale violence (Cilliers, 2018:22). There is thus a necessity for methods of national conflict resolution that interject at the transitional phase of conflict to break these patterns and address the root causes of such conflicts, to create longer lasting and sustainable peace.
Two countries in which the threat of recurrent conflict was (and to some degree remains) all too real, are South Africa and Rwanda. Both have histories of deep conflict and division, with the majority of South Africans having experienced extreme oppression under the apartheid regime and the Tutsi population having survived a genocide aimed at eliminating their entire nation group. It can be argued that both the apartheid regime and the Rwandan genocide were caused by and served to further exacerbate deep divisions within the citizenries of the two countries. Once these violent regimes came to an end, the broken nations left in their wake required mechanisms for conflict resolution targeted at the transitional era and the need for national reconciliation moving forward. The two mechanisms that resulted, namely the South African Truth and Reconciliation Commission (hereafter referred to as the TRC) and the Rwandan gacaca courts system, have come to be widely respected as successful projects in the resolution of deep-rooted conflict.
Despite not having been developed specifically as mediation efforts, these two projects demonstrate certain basic characteristics of the mediation technique. Both encouraged direct interaction between disputants1 in the form of a ‘truth-telling’ process, facilitated by a mediating body in the form of an institutional committee. Although these projects had differing primary objectives and practices, both aimed to mediate the expression of each sides’ experience so as to encourage reconciliatory mutual understanding between the disputing parties by the conclusion of the procedure.
Although the mechanisms displayed the basic characteristics of a mediation project, one would be hard-pressed to find a discussion of either in most ‘traditional’ discussions of mediation as
1 In both cases the disputants involved were perpetrators and victims of violent crimes, thus their interaction can
2 a tool for conflict resolution. These traditional discussions are most often focused on elite-level interaction, with the signing of peace agreements regarded as the most important indicator of mediation success. In developing a reconceptualization of the mediation project, and what constitutes mediation success, these two community-based African examples bring traditional conceptions into question. As opposed to the traditional focus on short-term mediation outcomes, these cases suggest that mediation could in fact be a method for creating longer-lasting peace in deeply divided African societies2.
1.2 Background and Rationale of the Study
As regimes rise and fall and power changes hands, conflict remains constant. Whether it be between individuals; ethnic/racial or religious groups; regions or countries, conflict and the divisions it creates affects every level of human and particularly political life. As such, the study of conflicts and their peaceful resolution has formed a constant, integral part of the study of society and politics, as scholars attempt to understand why conflicts occur and how best to resolve them.
The broad field of conflict resolution encompasses many different theories and mechanisms in attempts to tackle the issue of reconciling conflicted parties to one another. One such mechanism is that of mediation. In its most broad sense one could define mediation as “a process of dialogue and negotiation in which a third party assists two or more disputant parties, with their consent, to prevent, manage or resolve a conflict without resort to force,” (Nathan, 2012:1). As is the case with virtually any concept in the social sciences, any definition of mediation can undoubtedly be disputed, however the key component, that of a mediating body or institution facilitating discussion between disputants, remains the core of this approach to conflict resolution (Tillett & French, 2010:105; Fisher, 2011:159; Kressel, 2006:726; Touval & Zartman, 1985:7).
However, the preoccupation in general conflict literature with issues surrounding armed conflict and so-called ‘hot’ war has resulted in a parallel preoccupation within both the literature on conflict resolution and mediation itself towards the resolution of these kinds of conflict (Tillett & French, 2010:2). Being a mechanism that facilitates discussion between active disputants that would otherwise not have been likely to proceed (Zartman, 2001:8),
2 This study uses a characterisation presented by Guelke (2012:30), which regards a deeply divided society as
distinguishable from other societies by the existence of an entrenched fault line that is recurrent and endemic, containing the potential for violence between the fractured segments.
3 mediation has in traditional conceptualisations been firmly situated as a tool for conflict management or settlement (Mitchell, 1981:275; Burton, 1988:2; Ramsbotham et al, 2016:24). In other words, its utility has primarily been discussed in terms of bringing the leaders of disputing parties together and reaching some kind of agreement to reduce active combat. Throughout much of the mediation literature this assumption has permeated analyses, particularly with regards to measuring the success of mediation attempts, which is most often quantified by the signing of a peace agreement, settlement or ceasefire at their culmination (Touval & Zartman, 1985:14; Bercovitch & Gartner, 2006). With no regard for the longevity of such agreements or their contribution to building long-term, sustainable peace in deeply divided societies, this traditional practice and analysis of mediation arguably limits its utility as a tool for holistic conflict transformation. It is here that this study finds its point of entry into the scholarly discussion of mediation.
When situated as a tool for conflict settlement, the mediation mechanism is seen to be “task-oriented” (Fisher, 2011:159) and “interest-based” (Furlong, 2005:110). This outcomes-focused approach is therefore generally unconcerned with the nature of the disputants’ relationship and aimed more specifically at terminating immediate crisis situations and avoiding escalation of harmful violence (Wilkenfeld, Young, Asal & Quinn, 2003:282). However, this study will contribute to a trend of departure in mediation theory from regarding the tool as one for conflict settlement, to a tool for holistic conflict transformation. Transformation initiatives target the structural and institutional foundations of the conflict and address its cultural implications in terms of altering the way actors perceive themselves and their opponents in conflict situations (Rigby, 2006:47). By addressing these foundational elements of conflict and the relationships between disputants, mediation initiatives could have longer lasting results and create more durable, society-wide peace.
This study advocates that the key to viewing mediation as a tool for conflict transformation, as opposed to settlement, is moving from an ‘outcomes-based’ to a ‘process-based’ approach.3
This shift requires focusing primarily on the relational process of mediation and how the interaction between disputants can be used as a reconciliatory end in itself, as opposed to a
3 It must be recognised here that Horowitz (2007:57-58) does make mention of ‘process-focused mediation’ as a
trend in mediation practice. According to Horowitz, process-focused mediation limits the mediator’s role to that of ‘traffic lights’, concerned only with the physical process of the disputants’ dialogue, acting as a kind of negotiation administrator. This is the only place where mention of a process-oriented mediation trend has been found. For the purposes of this study however, the term ‘process-based mediation’ will refer to the model of mediation as developed in Chapter Two, focused on altering disputant relationships.
4 means to settlement ends. In this approach the mediators’ responsibility is no longer simply to facilitate agreements between disputing parties but is redirected to impact disputants’ relational interaction (Lederach, 2002:92). In this way the mediation effort could address the more deep-rooted conflict that exists between disputants and actively attempt to prevent exacerbation, recurrence or catalysing of overt tensions and potential violence.
1.3 Problem Statement and Research Questions
Mediation has been limited to a tool for conflict settlement in traditional theory and practice, with this trend extending into African mediation scholarship and practice. The assumption of this limited utility is however problematic in that it underlies all mediation efforts undertaken by key actors, shaping action taken by major institutions for conflict resolution such as the African Union Mediation Support Unit (AUMSU). As a result, all of the resources dedicated to mediation efforts by bodies such as the United Nations (UN), African Union (AU) and donor countries are directed at elite-level mediation operations that often prove unsuccessful or produce short-lived outcomes (Gartner & Bercovitch, 2006).
Although there has been some assertion of the need for documenting African mediation experience and drawing lessons from the past (AUMSU, 2016; Mottiar & van Jaarsveld, 2009; Govender & Ngandu, 2010), there remains much debate within the conflict resolution discipline over the link between the shortfalls of previous mediation attempts and the inadequacies of an outcomes-based approach to mediation. This study will aim to highlight these inadequacies and propose a shift in thinking with regards to the theory of mediation, suggesting that the potential of the mechanism lies not only in precipitating peace agreements, but in driving a process of society-wide conflict transformation. This could potentially inform mediation practice that shifts its focus from high-level meetings to community-based projects that are informed by the context in which they are conducted and address the divisions that both cause conflict and are left behind by conflict.
In addressing this research problem, the study will thus present the question: To what extent does ‘process-based’ mediation succeed as a tool for conflict transformation in African nations?
In order to measure the ‘success’ of African ‘process-based’ mediation projects in transforming conflict contexts, two sub-questions will frame the discussion:
- Under what conditions can a mediation project be deemed ‘process-based’? - Under what conditions can a mediation project be deemed ‘successful’?
5 1.4 Research Design and Methodology
This study will be done for descriptive and explanatory purposes. As the name suggests, descriptive research is done with the intention of describing phenomena and events. Descriptive studies, often qualitative in nature, answer questions of what, where, when and how (Babbie, 2011:68). This study will be descriptive in that the analysis will first undertake a description of the cases in question, delineating the historical context within which each took place, as well as the intricacies of the implementation of each project. Thus, the questions of what the projects constituted, where they were carried out, at which point in history they were undertaken, and how they were implemented will all be answered.
The study furthermore has a second purpose in explanation. Explanatory research is done to explain phenomena, or in other words to address questions of ‘why’ (Babbie, 2011:69). This study will aim to be explanatory in that it will explore why traditional outcomes-based mediation has thus far yielded mostly short-term solutions in African conflict resolution and propose a new way to utilise mediation in transforming conflict contexts and creating longer term peace.
1.4.1 Case Study Design
The research design of this project will be that of a case study design. A case study can broadly be described as “the in-depth examination of a single instance of some social phenomenon,” (Babbie, 2011: 301). The social phenomenon in this case being a mediation project that shows signs of being focused on the process of the mediation rather than its particular outcomes. According to Zartman (2005:3-4), case studies are arguably the most frequently used tool in studies concerning political negotiations and follow the trend of mediation research in focusing on the basic question of how outcomes are obtained through a negotiation. This is interesting in the context of the present research which challenges this focus of the mediation discipline, as in highlighting the process-related elements of the mediation projects the study will now also challenge the focus of the most commonly used research design within the study of political negotiations.
Kaarbo and Beasley (1999:372) provide a broad definition of a ‘case study’, stating it to be “a method of obtaining a “case”4 or a number of “cases” through an empirical examination of a
real-world phenomenon within its naturally occurring context, without directly manipulating
6 either the phenomenon or the context.” This definition is illuminating for this study as it identifies the design quite precisely but does not assume a specific purpose for this kind of investigation, as there can be more than one. Furthermore, it highlights the importance of context and lack of researcher’s direct manipulation of the phenomenon or this context (Kaarbo & Beasley, 1999:373).
The case study allows for understanding of phenomena in both their complexity and natural context (Miller & Brewer, 2003:22), with its main strength being that it provides in-depth insights into phenomena of potentially high complexity (Mouton, 2001:150). This design is therefore well-suited to the analysis of mediation mechanisms, which are most often employed in conflicts with high levels of content complexity (Tillett & French, 2010:139). While engaging with units of analysis such as nations, religions, social uprisings or, as in the case of this study, social projects, case studies help to develop intricate, cumulative historical explanations of complex cases (Bennett, 2004:19; Zartman, 2005:5). This type of case study can be referred to as ‘configurative-idiographic’, a design that attempts to build a near-total picture of the case, displaying the facts and not concerned with theory (Eckstein, 1992:136-138). In this type the purpose of the case study is description (Kaarbo & Beasley, 1999:373; Brown, 2018:16).
Given the descriptive purpose of this study, it will aim to describe the historical progression of the two projects as they took place in their natural context, particularly with regards to the characteristics of a ‘process-based’ mediation model. However, as stated by Brown (2018:20): “Deploying cases in this manner [for description] allows for the greatest possibility of authenticity and the least amount of generalisability.” This is due to the fact that no two practical cases are exact replicas, meaning that inferences drawn from in-depth analysis of a particular case cannot simply be generalised to others. This study, like most case studies, aims to extend beyond these limitations of purely factual description into the realm of theory-building and thus must extend beyond the bounds of idiographic research.
The configurative-idiographic type of case study is expressly not theory-driven. These studies are not guided by established or hypothesised generalisations, nor do they aim to generate theoretical propositions (Lijphart, 1971:691). However, this does not exclude the case study design from the scientific project of theory-building. Besides creating a clear historical picture of the cases in question, case studies may contribute to theory either by generating new hypotheses in a logic of discovery, or by testing existing ones in a logic of confirmation
7 (Bennett, 2004:21). The link of case studies to theory is said to operate on a continuum, in which the first stage sees theory being used to explore cases5, the next seeing cases used to
develop theory, in the next cases may be used to explore and refine theory, and finally cases may be used as tests of theory (Kaarbo & Beasley, 1999:374-376). The link to theory of the case studies in this research will be to explore and refine the theory of mediation, looking towards an updated model that could be tested further.
Case studies provide both wide and in-depth analysis of the cases in question and thus are able to model complex relationships. The importance of the descriptive ability of case studies is extended here, as it is the researcher’s in-depth understanding of a particular case that allows them to give context and meaning to data collected for causal explanations (Zartman, 2005:6). The complexity these models are able to encompass is a strength of the design when compared to the more straightforward, clinical relationships deduced by other approaches (Bennett, 2004:19).
Modelling complex causal relationships within specific case studies is therefore insightful but lacks generalisability as case study findings are often contingent on the specific conditions of the cases studied (Bennett, 2004:19)6. While it is logical that cases would have ‘within-case’
value, meaning conclusions will have validity in the context in which the study was done, it is not as widely accepted that any findings may be generalised beyond this context (Brown, 2018:18; Flick, 2009:134). The construct validity that allows for detailed analysis of contextual variables comes at the cost of external validity – producing generalisations applicable beyond the cases in question (Bennett, 2004:34). In this light the historical specificity that can be viewed as a strength, is seen as a weakness of the design and has led to a stereotype of case studies as atheoretical (Kaarbo & Beasley, 1999:371).
However, as exemplified by George & Bennett’s (2004:5) definition of a case study as “the detailed examination of an aspect of a historical episode to develop or test historical explanations that may be generalizable to other events,” there are scholars who believe insights from case studies may be extended beyond the specific cases analysed. Due to their high level of construct validity, case studies allow for contextualised comparisons of phenomena that may manifest differently in different contexts but are ‘analytically equivalent’ (Bennett, 2004:34; George & Bennett, 2004:19). It is for this reason that case study researchers often turn to a
5 This type of case study is defined by Eckstein (1992:138-139) as ‘Disciplined-Configurative’ – existing general
hypotheses can inform case explanations, but cases cannot in turn build or shape theories.
8 comparative methodology to harness the strength of highly contextualised conceptual understandings to compare ‘analytically equivalent’ phenomena and contribute to a theory-building project. This therefore supports the choice of methodology for this study, that of comparative analysis.
1.4.2 Comparative Analysis Methodology
Given the inherent limitations of case study research in building generalizable theory, an approach often used to offset some of the risk of overgeneralisation is to compare more than one case. As stated by Zartman (2005:7), “single case studies are of inherently limited utility in producing knowledge about negotiation as opposed to data on the unique case,” (emphasis added) and therefore in order to contribute to general theory-building, a comparative approach is useful.
Eckstein (1992:125) provides a definition of comparative case study commensurate with the approach to be taken in this research, defining it as “the study of numerous cases along the same lines, with a view to reporting and interpreting numerous measures on the same variables of different ‘individuals’.”7 He states further that by using a comparative methodology in
conjunction with case studies, theory may be successively refined through a ‘building-block’ technique in which regularities may be found through the simultaneous investigation of several cases (Eckstein, 1992:144). Miller & Brewer (2003:33) concur, stating comparative studies look for patterns of convergence between a single social process in different nation states. These patterns of convergence are intended to test and support the validity of a general theory that has been applied to understand and explain this social process.
Expanding on this definition, the historical comparative methodology constitutes “The examination of societies (or other social units) over time and in comparison with one another,” (Babbie, 2011: 347) and focuses on the similarities and differences between units of analysis in order to discover causal factors affecting an outcome seen in the set of cases (Mouton, 2001:154; Neuman, 2006:471). As stated by Mahoney and Rueschemeyer (2003:13), comparative historical analysis “makes possible a dialogue between theory and evidence of an intensity that is rare in quantitative social research.” The qualitative comparison of these two cross-national cases of mediation aims to give validity to a newer theory of mediation and thereby influence the construction of future projects using it as a tool for conflict
9 transformation. It is here however, in introducing the potential for causal analysis, that the method of comparing complex cases encounters a challenge.
Although the detail and complexity of case research provides in-depth understanding of case and context, when comparing cases to one another, comparisons must battle the weakness of having many variables and a small number of cases (Lijphart, 1971:685). If cases are used for what Skocpol and Somers (1980:181) define as ‘Macro-Causal Analysis’, controlled comparisons are intended to follow an experimental logic and provide valid causal inferences through multivariate analysis. However, with a multitude of potential extraneous variables and the practical impossibility of finding highly like cases in the real world, there is a definite danger of conclusions providing spurious relationships.
Skocpol and Somers (1980:176) argue however that the logic of ‘Macro-causal Analysis’ is not the only type of comparative analysis available to comparative scholars. They define two others, that of the parallel demonstration of theory and the contrast of contexts. Contrast-oriented comparatists, similarly to those conducting configurative-idiographic case studies, are focused primarily on the unique characteristics of individual cases and use comparisons to show how these characteristics influence general social processes within these specific contexts (Skocpol & Somers, 1980:178). Parallel-theorists on the other hand use the juxtaposition of cases to “persuade the reader that a given, explicitly delineated hypothesis or theory can repeatedly demonstrate its fruitfulness – its ability to convincingly order the evidence – when applied to a series of relevant historical trajectories,” (Skocpol & Somers, 1980:176). The similarity between cases under this approach is not necessarily in identical dependent or independent variable values (as would be necessary for strong causal conclusions) but lies in the common applicability of overall theoretical arguments. In this way these studies aim to show the validity of their theoretical arguments.
The present research will fit most comfortably into the latter approach, that of the parallel demonstration of theory. Due to the departure of this study from common mediation conceptualisations, and the resultant lack of similar cases based on this specific mediation model, the possibility of providing strong causal analyses based on multivariate analysis is limited. However, if a structured, focused comparative mechanism (as will be discussed below) can be produced from the theory as discussed in Chapter Two, this newer theory of mediation can be applied to cases with similar basic characteristics in an effort to determine the level of
10 applicability (and thereby the validity) of the overall theoretical arguments. This is how the study will fit the purpose of the Parallel-comparatists.
1.4.3 Alexander George’s Structured, Focused Comparison
In terms of strictly devised comparative methodologies, John Stuart Mill (1843/1967) is regarded as the father of comparative case study analysis, particularly with regards to his “method of agreement”8 and “method of difference”9. Following the logic of experimental
studies, these methods compare values of specific dependent and independent variables, with all other conditions kept constant. However, as has been stated, the reality in case study research is an impossibility of perfectly controlled case comparisons (Bennett, 2004:20). In using cases which may be influenced by different cultures, geographic locations, historical timing etc., controlling for every external variable becomes practically impossible.
One scholar who attempted to respond to these limitations of Mill’s methods is Alexander George and his method of structured, focused comparison of cases (George & Bennett, 2004:67-70). This study aims to achieve reliability and validity10 through this method which systematises the study of historical experience (George & Bennett, 2004:67) in a bid to develop more reliable and thus reproduceable case comparisons. Structured, focused comparison is regarded as the most-used approach for contemporary scholars engaging in case study research (Kaarbo & Beasley, 1999:370), and as the most recent authoritative statement thereof (Zartman, 2005:7).
This method is ‘structured’ in that it asks a set of standardised general questions reflecting the research objectives to guide data collection, making possible the systematic cumulation and comparison of case findings. It is also ‘focused’ in that it only deals with certain specific aspects of the historical cases under examination (George & Bennett, 2004:67). By using this set of questions and focusing on clearly identified classes of events, data can be systematically collected and compared (George & Bennett, 2004:69). Qualitative research in general is often critiqued for its lack of rigor, replicability and a systematic approach, thus this method provides a response to these problems (Kachuyevski & Samuel, 2018:3). It is when used in this way as
8 The researcher compares cases with the same value of the dependent variable to find potentially causal
antecedent conditions (independent variables) that are also the same (Bennett, 2004:30).
9 The researcher compares cases with different outcomes (dependent variables) to find antecedent conditions
(independent variables) that differ, judging that antecedent conditions that were the same despite the different outcomes were not sufficient to cause these outcomes (Bennett, 2004:31).
10 Reliability understood as the reproducibility of the research project, as opposed to validity understood as a
11 the basis for a systematic methodological approach, that case studies may form the foundations for analytical claims with external validity (Brown, 2018:16).
For the purposes of this study, George’s method of structured, focused comparison provides a helpful basic outline to follow. Using the mediation literature surveyed in the following chapter, a model for mediation projects to be conducted in deeply divided societies will be outlined, based on selected structured characteristics. The ‘questions’ thus guiding data collection will be to what extent each case demonstrates these characteristics. The comparison of cases can then be focused by this structure, with the analysis centring on the manifestation of the characteristics and how the transformative success of each project was impacted as a result.
By focusing the comparison on specific theory-derived elements of the cases in question, this approach aims to combat the problem of comparing highly complex cases (Flick, 2009:135). The approach can be discussed as a kind of ‘coding’ of qualitative data, in which themes and categories are pre-decided and used to structure both the data collection process and the case comparisons. This process can also be referred to as qualitative content analysis, a model of text analysis in which the basic goal is to reduce a multiplicity of available information (Flick, 2009:323).
1.4.4 Case Selection
The cases that have been selected as potential examples of ‘process-based’ mediation in Africa are that of the South African TRC established in 1995 and the Rwandan gacaca courts system founded in 2002.
The TRC and the gacaca courts have been studied extensively as mechanisms of transitional justice, particularly in terms of their significance in formal legal systems. However, they have yet to be included in discussions of mediation, particularly in the African context. Both cases show characteristics of community-style mediation between perpetrators and victims, with institutional structures facilitating dialogue between the disputants, and respected individuals forming the panel of mediators conducting these dialogue processes.
As this study is introducing its own theory-driven conceptualisation of mediation as a national conflict response, there are no well-documented cases of this specific type of mediation project. The cases selected therefore needed to be of a high-enough profile that there is sufficient secondary data available to build a complex case, but still achieve the comparative case
12 requirements of salience in the general discourse of African conflict resolution and relevance to the conceptual issues addressed in the study (Zartman, 2005:7). The comparison relies on the applicability of this mediation theory to the cases in question, thus they must exhibit the basic characteristics of a mediation project in order to be comparable at all, but vary on the values of each element of the newer model, so as to make deductions about the effects of each element on the overall success of the project at transforming the root causes of conflict. The specific selection criteria for these cases are based on the primary elements of the study’s conceptualisation of a process-based mediation project. This conceptualisation is broken down into two focal criteria. The first as has already been mentioned is the presence of a third-party mediator or mediating body, facilitating and encouraging discussion between disputants that otherwise would not have taken place (Zartman, 2001:8). These conflicts were suited to the mediation process, in that a third-party was needed to break down the barriers of negative feelings between disputants, a strong disagreement or difficulty to ascertain the ‘truth’ of the situation, and an absence beforehand of a negotiating forum (Kressel, 2006:732).
The second criterion, as the study is aimed at providing support for mediation aimed at the transformation of national conflict, is that the projects must have aims based on reconciliation, durable peace, and restorative justice. It is only with aims based on these key themes that mediation projects will focus on the change in quality of relational interaction between disputants that this newer theory of mediation espouses. Transformation extends beyond the immediate cessation of violence and creation of issue-based solutions (Mitchell, 2002:3), and endeavours to engender structural change to ensure long-term positive peace with social justice (Fisher, 2011:158). It would be impossible to assess conflict resolution projects on their ability to engender widespread conflict transformation if this is not what they were intended for. This is why the cases will not include any former ‘traditional’ mediation attempts in active African conflict that were primarily focused on the signing of peace agreements or ceasefires. The two selected mechanisms did have different goals and procedures, with the TRC being aimed largely at truth-telling and the granting of amnesties (Wielenga, 2011), while the gacaca courts had more of a mixture of aims between truth-telling, reintegration of perpetrators and in some cases a degree of punitive measures (Bowd, 2010). However, both projects had the overall aim of reconciling citizens in the transitional phase of their post-conflict societies and providing some measure of restorative justice. Thus, they both provide good examples of
13 mediation mechanisms aimed at conflict transformation, however with varying elements in their procedures and aims to be compared with the study’s model of ‘process-based’ mediation.
1.4.5 Data Collection
With improvements in technology the popularity of statistical methods boomed from the 1960s and 1970s, leading to increasingly negative assumptions about the capabilities of qualitative research (George & Bennett, 2004:6). Researchers in the social sciences have attempted to use technical statistical data to add credibility to their work in the scientific research community. However, as argued by Kachuyevski and Samuel (2018:7-8), highly technical research methods can in some cases create a divide between theory and practice as they may be overly lengthy and difficult for citizens or practitioners to fully appreciate. On the other hand, while quantitative methods provide stringent statistical interpretations based off so-called proxy indicators, qualitative methods allow for the interpretation of themes and patterns (Creswell, 2014:17), giving real-world context and actionable knowledge to political research (Kachuyevski & Samuel, 2018:8).
Therefore, the data to be used in this comparative analysis will be qualitative in nature. According to Babbie (2011:24), “Qualitative data are richer in meaning and detail than are quantified data,” and thus will be useful in building a detailed and in-depth understanding of the cases in question. Context is of great importance in the collection and analysis of qualitative data as there is a recognition of the complex intertwining between social processes and their social context (Miller & Brewer, 2003:193). This recognition of the importance of context is foreseen to be important in this study with its particular focus on African conflict. The African continent provides a unique social context that is inextricably linked to all social processes that take place therein, and thus must be examined in order to build the in-depth understanding a qualitative approach can provide.
As is typical in qualitative research, through an emergent data collection process this study will make use of various forms of data (Creswell, 2014:186), including secondary sources such as published books and journals, and primary sources in the form of organisational/government reports. This qualitative data will be collected by means of desktop research. The data will be organised or ‘coded’ according to the categories described above. As these projects began and culminated several years ago, it will be necessary to look at what they were intended to achieve at their inception, what they were able to achieve by their conclusion, and the changing perception of their impact on the respective nations. Particularly as the study aims to assess the
14 longer-term durability of conflict transformation resulting from these projects, it will be necessary to see how these perceptions have shifted over time. This will be achieved by assessing the literature produced concerning the projects and available data on public perception, such as that found in the South African Reconciliation Barometer (SARB)11 and the Rwandan Reconciliation Barometer (RRB)12.
1.4.6 Limitations and Challenges
As discussed by Mouton (2001:155), a limitation associated with the comparative methodology are problems with selecting “appropriate” cases that both fit within the selection criteria and are comparable to a high degree. As has been stated, this is a problem for this study in particular due to the lack of previous mediation attempts based on this model. However, to mitigate this limitation the cases selected both exhibit the basic characteristics of a mediation project and thus are comparable but vary on their individual demonstrations of the model’s characteristics. This will thereby make possible deductions concerning their varying successes. Mouton (2001:155) also notes that in cross-cultural and cross-national studies, as this one will be, there are constraints in comparability due to the differences in areas such as language, tradition and culture. However, in the context of this study these differences are in fact a necessary condition in order to evaluate the relevance of context-specificity and utilise this strength of the case study design.
A challenge facing this study in particular is the definition and measurement of ‘successful conflict transformation’. Definitively measuring the level of conflict transformation in a country proves difficult when the term itself encompasses such a vast array of conflict relationships. This study will therefore aim to build a broad picture of the level of conflict transformation in each country, based on triangulation of various sources of data concerning themes such as reconciliation, peace and restorative justice.
1.5 Contribution of the Study
By shifting the focus of mediation efforts from the outcomes it produces and towards the transformed conflict environment a mediation process could engender, this study could provide evidence for a new approach to mediation efforts. This approach could potentially provide
11 2017 SARB report produced by the Institute for Justice and Reconciliation, South Africa. The SARB uses
survey data to measure public perception of the state of reconciliation in South Africa.
12 Reports published in 2010 and 2015 by the National Unity and Reconciliation Commission of Rwanda in
collaboration with the United Nations Development Programme. Uses survey data to measure public perception of the state of reconciliation in Rwanda.
15 longer-lasting effects than those of the traditional outcomes-based approach. Furthermore, this study will focus on mediation in Africa, a continent with a long history of violent conflict that continues to this day and is in desperate need of new approaches to resolution. With peacebuilding resources being low in supply and high in demand on the continent, the conflict responses required are those that can efficiently and effectively address the deeper roots of conflict in a society and thus help to avoid lapses back into the costly and persistent cycles of violence. This study could provide evidence for mediation as a conflict response with the potential to achieve such goals, provided its practice and utility is reconceptualised as a process of conflict transformation and not simply agreements for conflict settlement.
In discussing and describing a contemporary holistic approach to mediation and comparing this to similar mediation attempts already having been carried out on the continent, this study would aim to make both a theoretical and practical contribution, thus being both basic and applied in nature. Theoretically, it could provide support for a newer model of mediation, moving away from the traditional outcomes-based approach. Practically, it could provide evidence that this model is one worth emulating in future peacebuilding efforts on the continent and give practitioners in the business of conflict resolution reason to revisit the conduct of mediation as a conflict response tool in transitional societies.
1.6 Outline of the Study
This study will be presented in five chapters. This first chapter has served as an introduction and outline of the study to be conducted, including the research design and methodology. Chapter Two will form the conceptual and theoretical framework of the study with a discussion of traditional mediation theory and the major debates surrounding its description as a mechanism for conflict management, as well as mapping the progression of mediation theory towards the goal of broader conflict transformation. This chapter will further provide the analytical framework for the study in the form of a model for process-based mediation projects. Chapters Three and Four will then be dedicated to the construction of the cases and the analysis thereof with regards to the analytical framework produced in Chapter Two. Chapter Three will be dedicated to the historical contextualisation of each case, with a discussion of the conflict climate in each nation that necessitated the resolution projects, followed by a description of these projects. The data collection and presentation for this study will be done in Chapter Four, where the comparative analysis of the TRC and the gacaca courts will take place. Following the historical description of the cases, the theory of ‘process-based’ mediation will be overlaid
16 on to the cases, highlighting their relative manifestations of the criteria included in the study’s conceptualisation of this new type of mediation project. The two cases will be contrasted and compared, looking both at how they are similar and how they differ in demonstrating a process-based project, and how these differences have led to the relative conflict transformation of each historically conflicted nation.
Finally, Chapter Five will follow with a discussion of the findings in Chapter Four and relate them back to the research question as originally posed in Chapter One. This chapter will serve as a conclusion to the study and provide recommendations for further research.
17 CHAPTER TWO – CONCEPTUAL, THEORETICAL AND ANALYTICAL FRAMEWORK
2.1 Introduction
Just as conflict has retained a constant presence throughout human history, so has mediation played an integral part in its resolution. Evidence of mediation practice is found in cultures throughout the world, and in ages as far removed as the Biblical – from Moses acting as mediator between God and men to shamans in Asia, America and Oceania mediating with spirits (Horowitz, 2007:51). However, in the contemporary era with weapons so destructive and technology so developed, mediation has become an even more important tool in terminating violent conflict that is both costly and irrational (Bercovitch, 1996:2).
The upsurge in mediation practice, particularly in the post-Cold War era, has led to a consequent surge in scholarly work surrounding the discipline, with many adding their analyses to the conceptualisation of the term ‘mediation’. Upon examination of these varied analyses it becomes clear that there are at least two broad approaches to the practice, a traditional and a non-traditional. The traditional can be described as a ‘problem-solving’ approach, situated within the field of conflict management. The non-traditional however is more of a ‘transformative’ approach, aptly named due to its situation within the field of conflict transformation. Although this study is concerned with the potential of mediation as a tool for conflict transformation, in order to provide full reasoning for this preferred conceptualisation, it is necessary to provide an overview of mediation in all its varied forms; the aim of this chapter.
2.2 Mediation Defined
As one of the primary mechanisms utilised and discussed in the field of conflict resolution, the concept and practice of ‘mediation’ are subject to much debate. In structuring an initial conceptualisation of the term, an often-cited definition of the term ‘mediation’ provided by Christopher Moore is useful. Moore (1986:14) describes mediation as an “intervention into a dispute or negotiation by an acceptable, impartial and neutral third party, who has no
authoritative decision-making power to assist disputing parties in voluntarily reaching their
own mutually acceptable settlement of issues in dispute” (emphasis added).
This definition, while including every aspect of mediation practice that are often assumed integral, in fact provides a concise summary of its most contested elements. In order to
18 contribute to the discussion of mediation best practice, as this study aims to do, it is important to begin with an analysis of these key elements and the various debates surrounding them.
2.2.1 Conflicts Conducive to Mediation Intervention
As a result of overwhelmed legal systems and the often exorbitant costs associated with legal disputes, mediation has increasingly been utilised as a form of Alternative Dispute Resolution that relieves courts and can reduce costs of resolving conflict (Fisher, 2011:159). Despite its popularity and widespread practice however, it must be recognised that mediation is but one of a plethora of conflict resolution mechanisms. Furlong (2005:124) identifies a key element that is often regarded as the signifier for mediator intervention as opposed to any of these other mechanisms, that of the stalemate. He states that prior to parties reaching a stalemate, pure negotiation through direct interaction between disputants is the tool most likely to be implemented. However, there are other signifiers that suggest mediation to be the most appropriate approach and these will be discussed below.
Kressel (2006:732) distinguishes three sets of barriers which may impede pure negotiation and thus encourage the intervention of a mediator. The first of these are interpersonal barriers, in which negative feelings between disputants or dysfunctional communication between them infringe on the quality of direct interaction. The second are substantive barriers, in which disputants disagree strongly on the content of the disagreement and have difficulty ascertaining the ‘truth’ or ‘facts’ of the situation. The third set of barriers are procedural, in which there is an absence entirely of a negotiating forum and thus a complete deadlock. Mediation may break down these barriers as it provides an opportunity for constructive communication in which all aspects of the conflict and possible solutions can be fully investigated and weighed up, potentially opening new political space that previously had not been seen (Ramsbotham, Woodhouse & Miall, 2016:213).
Tillett & French (2010:139) ascertain a set of pre-conditions for mediation to take place, these being: previous failure to resolve the conflict through collaboration, disputants’ perception that they will not be able to resolve their issues alone, distrust or hostility between the disputants, extended duration of conflict, and a high level of complexity in the content of the conflict. Similarly, Bercovitch and Houston (1996:12) suggest mediation is likely to be called upon when the conflict has gone on for some time, if previous efforts have ended in stalemate, if both actors face unendurable costs if the conflict were to continue, and if the parties to the conflict are open and willing to engage in mediation through direct or indirect dialogue.
19 Although mediation is intended to provide a level playing field for negotiation, inevitably in some cases there will be an imbalance of power between disputants. Weaker actors may therefore turn to mediation in the belief that they could achieve a better outcome than they would otherwise have done without the mediator present. Parties may also agree to the mediation in the hopes that the mediator will act as a guarantor for any agreements reached, reducing the risk of either side reneging on the agreement (Touval & Zartman, 1985:9). In the case of victim-offender mediation and traumatic conflict, the presence of a third-party may make the affected victim feel safer in attending the negotiation (Furlong, 2005:124).
Timing is also an important factor as contended by Zartman’s (2001:8) ‘ripeness theory’, in which mediation is considered the appropriate response once conflict has reached a
“mutually-hurting stalemate”. In other words, mediation should only take place when conflict is ‘ripe’,
and a cost-benefit analysis demonstrates two conditions: that neither party is willing nor able to make any more power moves towards winning the conflict, and that the continuation thereof has become detrimental to all parties involved.
In an empirical study concerning the effectiveness of mediation in mitigating international crises, Wilkenfeld, Young, Asal and Quinn (2003:282) found that mediation is two or three times more likely to be the conflict response of choice in times of crisis and hostility than in periods of lower intensity. This being the result of pressure due to a perceived deadline and the immediate consequences if an agreement fails to be reached. The study also found that the higher the number of relevant issues in a crisis, the higher the chance of the crisis being mediated (Wilkenfeld et al, 2003:285). Finally, it showed that territorial crises, conflicts involving geographically close disputing groups and those involving ethnic divisions, all of which have been the trend post-Cold War, are more likely to feature mediation than conflicts without these key characteristics (Wilkenfeld et al, 2003:286).
In contemporary warfare with the reality of nuclear arms and the extended duration of conflicts, mediation provides an alternative to the zero-sum game that is resolution of conflict through ‘winning’ or ‘losing’ a war. It is said to differ from other forms of third-party intervention in that it is an act of peacemaking, which in theory does not intend for one party to triumph over the other (Touval and Zartman, 1985:7). It is thus possible to resolve active conflict by reaching win/win outcomes in which the interests of all parties are considered and included in the decision-making process (Furlong, 2005:110). For disputants trapped in cycles of violence or halted in a stalemate, the push they may need is intervention from a mediator whose
20 fundamental objectives are to help them find a mutually acceptable solution and to offset their tendencies towards a competitive win/lose mindset (Kressel, 2006:726).
2.2.2 Bias, Neutrality and Impartiality
The next element for discussion, and possibly the most highly contested element of the mediation mechanism is that of the bias, neutrality or impartiality of the mediator themselves. While it is commonly agreed that mediators cannot be a party to the conflict in question (Fisher, 2011:161), there is a general assumption evident in many conceptualisations that mediators must be both neutral and impartial (Tillett & French, 2010:105; Fisher, 2011:159; Horowitz, 2007:51). In some texts, mediators are even referred to as “third-party neutrals” and in practice are often bound to the principle of neutrality by ethics codes (Wehr & Lederach, 1996:56). Tillett and French (2010:140) go as far as to say that a neutral third-party is the ‘distinguishing factor’ of mediation from similar processes such as collaborative decision-making and conflict resolution more broadly.
Within the North American scholarship in particular, the ideal mediator is expected to have both externality and neutrality, thereby coming from outside the conflict situation and having no connection to either side (Wehr & Lederach, 1996:57-58; Horowitz, 2007:53). The mediator’s role is defined negatively, focusing on what the mediator is not as opposed to what they are or could be. This includes mediators being unconnected to disputants, unbiased, having no investment in a particular settlement or outcome, and not expecting any reward from either side (Moore, 1986:15-16). Intuitively this may make sense, as in order to ensure a fair and equal process one would assume the mediator should have no reason to favour one side or the other. In the case that a mediator were to show bias towards either party, they may risk losing the trust of the other (Touval & Zartman, 1985:9).
In practice however, it is often the case that the third-party is partisan or allied to one side of the conflict. This may cause them to push solutions supporting that side or adding resources in its favour (Mitchell, 1981:275). For example, in the early 1990s at the beginning of the mediation of the Sudanese conflict, the mediating countries had different allegiances to the primary conflicting parties. Through the process this would change, as Eritrea, Ethiopia and Uganda came to support the Sudan People’s Liberation Army/National Democratic Alliance (SPLA/NDA) over the government of Sudan. With the support of these three mediators, by 1997-1998 the mediation had reached the point of agreement with a solution favouring their preferred side of the conflict, the SPLA/NDA (Svensson, 2009:448). However, contrary to the
21 assumption of the negative impact of bias and partiality of mediators, many authors argue these qualities could in fact be productive in the overall mediation process.
For example, based on their experience as mediators in Guatemala and Nicaragua, Wehr and Lederach (1996:58-59) introduce the conceptualisation of a ‘new’ kind of mediator role, that of the ‘Insider-Partial’. This mediator is said to be best suited to traditional societies with community-based norms and values. The ‘Insider-Partial’ is purposefully drawn from within the conflict context as their acceptability stems from extensive trust and connections with the conflicting parties. The success of this mediator relies on their proximity to the conflict, partly due to the fact that they do not leave the conflict setting once negotiations have ended, and from the accumulated knowledge of the conflict context shared by disputants and mediator (Wehr & Lederach, 1996:58). The parties are assured of the genuine vested interest the mediator would have in seeing the conflict effectively reconciled, and the mediator has extensive contextual knowledge to assist in achieving this goal.
Touval and Zartman (1985:8-9) argue it is in fact rare for mediators to be completely indifferent to the outcomes of the mediation, and that they almost always have some form of self-interested motivation for taking part. They list two such motivations, the first being that continued conflict may in some way disadvantage the mediator, for example if the conflict has gone unresolved for some time and the mediator wants to prove their problem-solving skills. The second is the potential to extend and increase their own influence through developing a stronger relationship with the parties involved. Touval and Zartman go on to state explicitly that impartiality is not important for successful mediation as parties accept the mediator based on their perceptions of what their own relationship with the mediator would be and the potential consequences of accepting or rejecting the eventual terms of agreement. They claim that parties are not as concerned as might be expected with the mediator’s general attitude and relationship towards them prior to the mediation (Touval & Zartman, 1985:15).
An author who attempted to empirically test these varying hypotheses on mediator bias is Svensson (2009). On the one hand his analyses of biased and neutral mediators in civil wars showed that biased mediators can positively and significantly affect the likelihood of specific institutional agreements within peace agreements. On the other they showed that because neutral mediators’ sole purpose is getting parties to come to an agreement, they are less concerned with the provisions set out in the agreement. Therefore, neutral mediators create outcomes “without provisions for political and territorial power-sharing, third-party security
